1. technology and computing
2. software
3. databases

MIS Review Vol. 20, No. 1, September (2014), pp. 31-43
DOI: 10.6131/MISR.2014.2001.02
© 2014 Department of Management Information Systems, College of Commerce
National Chengchi University & Airiti Press Inc.
Data Management Issues and Data Mining of Real Time
System Application for Environment Monitoring
Dinesh Kumar Saini1,2, Sanad Al Maskari3
Faculty of Computing and Information Technology, Sohar University, Oman
Research Fellow and Adjunct Faculty, School of ITEE, University of Queensland, Australia
3
School of ITEE, University of Queensland, Australia
1
2
ABSTRACT:
Environment pollution monitoring and control is very big problem for the whole
world. Taking decision in the environment is becoming more challenging. The aim
of this paper is to present the challenges surrounding environmental data sets and
to address these in order to develop solutions. Environmental data sets present
a number of data management challenges including data collection, integration,
quality and data mining. Environment data sets are also very dynamic and this
presents additional challenges ranging from data gathering to data integration,
particularly as these data sets are normally very large and expanding continuously.
Statistical methods are very effective and economical way to analyze small, static
data sets but they are not applicable for dynamic, real-time and large data sets.
The use of data mining methods to discover hidden knowledge in large datasets
therefore presents great potential to improve environmental management decisions.
A representative environmental data set from quantitative air quality monitoring
instruments has been assessed and will be used to demonstrate some of the issues
in applying data mining approaches to poor data quality.
KEYWORDS: Data Management, Real Time Systems, Data Mining, Environment Monitoring
Systems.
1. Introduction
Huge amount of data are generated by environmental sensors every day across the
globe. There is tremendous need for data analysis systems which are able to mine massive
and continuous stream of real world data applications such as temperature monitoring,
air pollution, stock market, network security, etc. Data generated by environmental
sensors are recorded at time intervals of seconds through to minutes and over time these
sensors will create datasets that need to be mined in real time in a way that takes into
considerations the dynamic nature s of the real world changes that are being measured.
Without appropriate analysis methods that allow inferences to be derived based on
patterns observed within these data sets it will not be possible to lead to new knowledge
discoveries (Fayyad, Piatetsky-Shapiro & Smyth, 2006) defines knowledge Discovery in
06-Saini.indd 31
2015/4/1 下午 02:07:36
32 Dinesh Kumar Saini, Sanad Al Maskari
Databases as “the nontrivial process of identifying valid, novel, potentially useful, and
ultimately understandable patterns in data.” In this research we are attempting to identify
best possible data mining solutions for environmental data using various algorithms,
focusing on streaming data mining methods as these types of data are increasingly
common now as a result of rapidly emerging technology developments. In order to do
this we have developed an environmental data mining model (EDMM) which covers all
critical stages needed to produce high quality data mining results. A key tenant of the
model is that data analysis methods must take into consideration the dynamic behaviors of
the incoming data streams. This means that statistical methods are not always appropriate
for the large, dynamic data sets characterizing many environmental variables. The
dynamic nature of such data means that their behaviors change over time and can lead to
concept drift. The occurrence of concept drift (Helmbold & Long, 1994) will hinder the
accuracy of the original data model thus make it increasingly inaccurate over time. In our
model we will take into consideration the effects of concept drift to reflect provide greater
data model accuracy.
Air pollution, marine pollution, and sand storms are three major problems that affect
agriculture and fishing, which need to be addressed and treated proactively (Al-Maskari,
Saini & Omar, 2010). Air pollution is a very critical problem that can lead to catastrophic
outcomes if not managed and monitored proactively.
2. Research problem
We took as case of Oman at Sohar Port for the current study of our research. The
new born industrial zone introduced a mixture of chemical and petrochemical plants that
leads to odorous emissions. Such emissions in a populated area cause nuisance discomfort
to citizens, and health issues. The government has installed a network of sensors mobile
and static to monitor odorous emissions in the industrial zone. The electronic sensors are
provided by Common Invent, Delft, The Netherlands All sensors communicate by wireless
or telephone links with a central database management system maintained by Common
Invent which handles over 2 million new data entries per day. The data base management
system reads and stores the incoming raw data, as well as providing some interpretation of the
data and coordinates automated event handling for clients (see http://www.comon-invent.com).
Electronic nose (or e-nose) is the name given to a wide class of instruments capable
of measuring odor information in different environments (Saini & Yousif, 2013). The
data from the e-noses provided by Comon Invent has been used to distinguish odorous
petrochemical vapors, like fuel oil, naphtha, gasoline, jet fuel, from manure, from sewage
gas or from VOCs like toluene, benzene, styrene etc. (Bootsma et al., 2012). In Sohar Port
06-Saini.indd 32
2015/4/1 下午 02:07:36
Data Management Issues and Data Mining of Real Time System Application for Environment Monitoring 33
the objective is to develop the e-nose data assessment system to be trained to associate
digital fingerprints with various odors. Once trained, the e-nose database system can then
recognize smells as they arise and inform the user as to their origin once this has been
established. The current e-noses deployed at Sohar Port create a record every 3 minutes
and stores the raw data in a remote server for the eight different sensors located in each
unit. The system also records wind direction, speed, and air temperature, as well as GPS
location and time. Currently a total of 7 static sensors and one mobile sensor are installed
in the industrial zone with another two sensors installed in an adjacent urban community.
Consequently data collected from these distributed sensors forms a large data source,
which needs to be cleaned, mined and analyzed in different dimensions to create a usable
prediction models and for consumption by different applications.
Data gathered from different environmental sensors needs to be analyzed and
assessed appropriately. Any system that handles critical data that can reduce air quality
impacts on the adjacent communities has to be credible and effective. Credibility demands
that the system is accurate when dispatching a poor air quality alert.. The system will need
to monitor and predict odors emissions in a distributed area taking into accounts external
elements such as wind, direction and speed, rainfall and dust storms. The system has to be
smart enough to issue an alarm based on area of effect and the likely degree of impact. In
developing such a system the following are the main challenges faced in this research:
(1) Data integration and data quality issues need to be addressed before applying data
mining techniques.
(2) Multiple data sources from different stake holders that need to be integrated.
(3) Odor sensors data sets have limited finger prints which make it hard to create an
accurate prediction model.
(4) Current data mining methods used in environmental analysis do not take into
considerations the surrounding environment changes and their dynamic behaviors.
(5) Algorithm evaluation in the scope of environmental monitoring is complicated by the
complex nature of environment systems.
(6) Using statistical methods to analyze dynamic, real-time and large data sets are not
appropriate (Arasu et al., 2003).
3. Research approach
The combination of emerging new semantic web technologies and web services to
access, process and integrate data and models held within both centralized and distributed
hydrological databases allows:
06-Saini.indd 33
2015/4/1 下午 02:07:36
34 Dinesh Kumar Saini, Sanad Al Maskari
(1) Identification and prioritization of the key stakeholder user requirements, queries and
datasets.
(2) Data quality and data cleansing processes.
(3) Development of the common data models and ontologies to integrate both static and
real-time data streams, visual, spatial and temporal data, legacy databases and newly
generated datasets.
(4) Prediction of the Air pollution: Using the captured data we intend to create models to
identify the air pollution sources, delineate affected areas and estimate next areas to
be polluted. A learning system will be employed to enable a better sensing of odors.
Beside the data collected from sensors, feedback from experts, researchers, and
community will be assessed to see if these improve the predictions. A community
enabled sensor network will be considered to improve the odor prediction model.
(5) The data collected by the e-noses will be analyzed using various data mining and
machine learning techniques. A prediction model for the air pollution effects will be
the core focus of research. In the prediction model we will take into consideration the
metrological variables (temperature, wind speed, wind direction, and rain). Clustering,
Euclidean, Cosine similarities, ANN, CVFDT, Fuzzy logic and other techniques will
be used and compared to achieve best prediction model.
(6) The output of the prediction model will be evaluated against criteria developed for an
industry and community alert system.
4. Motivation
Information technology has to be proven to be cost effective and provide a reliable
solution if it is to be deployed effectively in environmental management prevention and
control (Abdelzaher et al., 2010). The deployment of distributed network sensor has grown
rapidly in the past ten years (Hill et al., 2000). Many applications have been implemented
using network sensors such as environment monitoring, intrusion detection (Wood &
Stankovic, 2000), habitat monitoring (Crumiere, 1999), defense, transportation, and
scientific exploration. These sensors produce very large volumes of data that need to be
cleaned, stored, and retrieved for analysis and decision making (Li et al., 2000). The huge
amount of data produced by these sensors need to be monitored and controlled to provide
a meaningful and useful information and it can be used as a feedback loop to management
and control systems.
Environment monitoring applications can be very complex because they involve
many variables with different dimensions and different scales. Managing, accessing
06-Saini.indd 34
2015/4/1 下午 02:07:36
Data Management Issues and Data Mining of Real Time System Application for Environment Monitoring 35
and analyzing data generated from distributed environmental sensors remains a serious
challenge for researchers and scientists. The complex data set created by these sensors
present many challenges including visualization, data storage and retrieval, data quality
and data integrity as well as data analysis and mining.
The ability of sensor network to deliver large amount of data in real-time create
a data mining challenge. In our research we are attempting to cover these challenges
and provide solutions to such issues. Most data mining algorithm doesn’t take
into considerations the dynamic behaviors of multi dimensions data generated by
sensor networks especially for environmental data. Concept drift must be taken into
considerations when analyzing environmental data in real time.
5. Environmental network sensors
The new developments in the area of electronics and wireless network have led to
the creation of Environmental Network Sensors (ENS). ENS are large distributed systems
communicate through wireless, telephone or satellite networks to stream the data to a
central location for processing and analysis (Sohn et al., 2003). A fundamental aspect of
environmental network sensors deployment is therefore data management and analysis
of these data streams. To enable environmental organizations and authorities to be able
to make constructive decisions regarding likely environmental impacts environmental
informatics need to be in well shape and in place.
6. Why data mining
Traditional methods for analyzing data using statistical methods are limited to very
small data sets and to single users dealing with them directly. Although statistical methods
are very economical, simple and effective but they are complicated when trying to apply
them to new applications (Friedman, n.d.). Unlike data mining methods, they are not
meant to deal with huge, dynamic and real time data sets. Most statisticians will consider
1,000 point as a large data set but in data mining world millions of transactions can be
analyzed using data mining algorithms (Roppel et al., 1998).
The evolution of data mining techniques started with the advent of the first computer
devices within the university environment and more broadly with the subsequent
development of personal computing devices (Pan & Yang, 2007). The improvements
of data communication, networks, databases and the ability for users to ubiquitously
access data and services in real time has revolutionized the data mining field. Each new
technology and development of numerical techniques was based on the previous one. The
following shows how we have developed through the data mining era:
06-Saini.indd 35
2015/4/1 下午 02:07:36
36 Dinesh Kumar Saini, Sanad Al Maskari
(1) 1970s era: Data collection
In this era data was collected using computers, tapes and discs.
(2) 1980s & 1990s: Data Access
The introduction of Relational DBMS, Relational data was implemented, Data
RDBMS, advanced data models (extended-relational, object oriented [OO], deductive,
etc.) Application-oriented DBMS (spatial, scientific, engineering, etc.)
(3) 2000s: Data mining era
The introduction of the WWW and the maturity of the internet have created massive
databases. Data mining algorithms started to emerge and advanced data mining
algorithms has been introduced. New applications in multimedia, web mining,
streaming data management and mining were deployed and still apply (Crumiere,
1999).
Data mining is mature enough to be applied in environmental applications and other
large data applications due to:
(1) Massive data collected for instance large distributed sensors are used to collect large
and complex data about the nature and pollution.
(2) We can now access powerful multiprocessors and data storage technologies at
reasonable prices.
(3) Many data mining algorithms have been developed.
(4) Data mining is concerned with creating knowledge and information from dynamic and
huge data sets. It is a blended field of statistics, machine learning and data bases. Data
mining also referred to as Knowledge Discovery in Database (KDD). The definition
of data mining varies between different authors based on their own background,
experience and views. For example:
● Data mining is the process of exploration and analysis, by automatic or
semiautomatic means, of large quantities of data in order to discover meaningful
patterns and rules.
● Data mining is the process of extracting previously unknown, comprehensible and
actionable information from large database and using it to make crucial business
decisions.
● Data mining is finding interesting structure (patterns, statistical models,
relationships) in databases.
● Data mining is a decision support process where we look in large databases for
unknown and unexpected patterns of information.
06-Saini.indd 36
2015/4/1 下午 02:07:36
Data Management Issues and Data Mining of Real Time System Application for Environment Monitoring 37
The use of data mining techniques in e-nose odor monitoring is still in an early
stage. Few modern data mining techniques have been used in e-nose odor monitoring
and analysis. Artificial Neural Networks have been applied to predict Sulphur dioxide
concentration in Delhi (Green, Chan & Goubran, 2009), to process the signal from
odor sensor arrays for near-real-time odor identification (Kahn, Katz & Pister, 1999),
using electronic nose to identify the age of spoiled food based to predict piggery odor
concentrations (Foster, 2002). In his paper (Aberer et al., 2010) describes the vision of
community based sensing using a mobile geo-sensor network (Gehrke & Madden, 2004).
The OpenSense project is aiming to investigate air pollution monitoring using communitydriven sensing (European Commission, 2001).
7. Environmental data mining model
In this section we will describe our Environmental Data Mining Model (EDM)
that we will use to create a prediction model using various data mining algorithms. The
EDM has eight essential stages starting from data collection and ending up with decision
making and result monitoring (refer to Figure 1). The use of this model will optimize the
efficiency and accuracy of our data mining models.
7.1 Data collection
Data collection can be done manually or automatically using sensors. The Sohar
Environment Unit (SEU) uses Mobile Air Quality Monitoring Station (MAQMS) which
YES
Decision Making
Rules, Regulation
& Standards
!
NO
Result
Monitoring
Data Collection
Decision
Formation
Continual
Improvement!
Data
Mining
Data
Selection
Start
Data
Inspection
Data
Integration
Figure 1 EDDM Model
06-Saini.indd 37
2015/4/1 下午 02:07:36
38 Dinesh Kumar Saini, Sanad Al Maskari
record data hourly. MAQMS record PM10, O3, CO, SO2, H2S, NO, NOx concentrations
with the corresponding meteorological data. The new distributed network sensors (e-Nose)
records data every 3 minutes. The E-Nose data are designed to monitor odors around
Sohar Industrial Port (SIP). The reading from the E-Nose sensors doesn’t refer to any gas
instead it writes some values based on the sensor reaction (refer to Table 1).
7.2 Data inspection
Current data collected by SEU are not hosted in centralized warehouses rather it’s
saved in Excel spreadsheets. This present an issue with data cleansing and integration
between distributed sensors. Raw environmental data i is usually not clean, incomplete
(empty values), noisy (contains error) and may have inconsistencies (Al-Maskari et
al., 2010). The following table illustrates the multi source problems identified in SEU
environmental datasets:
The employees of the system have all the rights to expect that the data they are
dealing with are indeed correct. Otherwise wrong decisions will be made and the
consequences of such decisions can be significant to them and to the environment
Table 1 Data Quality and Cleaning Issues
Problem
Attribute Missing Values
06-Saini.indd 38
Dirty Data
O3 = “No Data”
Remarks
Value unavailable during
data gathering (null value)
Different values with same O3 = Zero, O3 = 0
meaning
Both refer to the same value
Record
Violated data type
O3 = Zero, O3 = Span
The value should be numeric
Record
Type
Duplicate records
Sensor 1 (Date = 1/7/08, Duplicate records from
time = 11:00, dust = 0.025) different data sources
Sensor 1 (Date = 1/7/08,
time = 11:00, dust = 0.025)
uniqueness violation
Sensor 1 (Date = 1/7/08, The primary key used in
time = 12:00, dust = 0.016) both sensors are date and
Sensor 1 (Date = 1/7/08, time
time = 12:00, dust = 0.025)
Different time format
Sensor 1 (Date = 29/02/08, Different date and time
time = 0:00)
format is used by different
Sensor 2 (Date = 02/29/08, sensors
time = 24:00)
Different unit formats
Sensor 1 Dust unit ug/m3
Sensor 2 Dust unit mg/m3
Different units are used by
different sensors
Noisy
e-nose S1 = 0, -0.12
Errors and outliers
2015/4/1 下午 02:07:36
Data Management Issues and Data Mining of Real Time System Application for Environment Monitoring 39
Poor data quality management of environmental data can lead to the following:
(1) Increase cost as more time will be spent correcting errors rather than performing
critical operations.
(2) Poor data quality may lead to poor decision making which lead to incorrect estimate
and predictions. E-environment is so sensitive that we should not tolerate any
compromise when it comes to decision making or we will endanger or people and
planet earth.
(3) More difficult to set strategy and execute it. Environmental strategic decision requires
data gathering from various data sources with some uncertain quality. This makes it
harder to develop a sound strategic decision. Executing the strategy becomes difficult
as inaccurate results become evident.
To overcome the above concerns the following operations will be conducted to the
data sets:
(1) Fill in missing values.
(2) Identify outliers and smooth out noisy data.
(3) Correct inconsistent data.
(4) Duplicate identification.
7.3 Data integration
Environmental data is heterogonous by nature therefore combining data from
different sources can be a very challenging task. Legacy data, heterogonous data, time
synchronization can be a source of problems when integrating data from different sources.
Before integrating multiple data sources they must pass data quality checks, otherwise
data quality problems will be inherited from the source databases. By looking at Sohar
industrial region environmental data we can observe the multi-source cleansing problem
(refer to Tables 1 and 2).
Table 2 Data Collection for Various Gases on the Same Day at Different Time
Date
Time
Dust
O3
CO
nCH4
SO2
mg/m
ppb
ppm
ppm
PPB
3
06-Saini.indd 39
01/07/2008
10:00
0.37
S<
S<
RS232
3.608
01/07/2008
11:00
0.025
36.75
0.86
RS232
1.249
01/08/2008
0:00
No Data
15.82
0
RS232
0.125
2015/4/1 下午 02:07:36
40 Dinesh Kumar Saini, Sanad Al Maskari
7.4 Data selection
Data inspection and integration is considered to be the most difficult and longest
stages in EDMM. After creating integrated clean data sets it’s necessary to define
application domain for the data mining algorithm. Meta data information, prior knowledge
and application goals must be defined. Once they are defined a target data set will be
generated. One of our objectives is to predict odors in the SIP area. We will use E-Nose
data combined with social network feedback dataset.
7.5 Data mining
In this stage an appropriate data mining approach will be selected. Various data
mining approaches exist including association rule mining, clustering, induction and
streaming data mining. Once the appropriate approach is selected then a suitable
implementation will be applied. Data mining can focus on a variety of areas.
Typical areas to examine are:
(1) habits and behavior,
(2) demographics,
(3) time,
(4) product characteristics.
8. Conclusion
In this paper we introduced our Environmental Data Mining Model which addressed
all aspects surrounding environmental datasets. The EDM introduced eight essential stages
necessary to create an accurate data mining results. EDM is a continuous improvement
model aiming to improve the process of environmental data mining. EDM will help in the
improvement of decision making and environment pollution control.
Acknowledgements
The authors of the paper are thankful for the University Research Council and Sohar
University, Oman. Authors are thankful for the University of Queensland, school of ITEE.
Thanks to Prof Lance Bode and Prof Tony for the Valuable Input.
06-Saini.indd 40
2015/4/1 下午 02:07:36
Data Management Issues and Data Mining of Real Time System Application for Environment Monitoring 41
References
Abdelzaher, T., Blum, B., Cao, Q., Evans, D., George, J., George, S., et al. (2010), ‘EnviroTrack:
towards an environmental computing paradigm for distributed sensor networks’,
Proceedings of the 24th International Conference on Distributed Computing Systems,
Tokyo, Japan, pp. 582-589.
Aberer, K., Sathe, S., Chakraborty, D., Martinoli, A., Barrenetxea, G., Faltings, B., et al. (2010),
‘OpenSense: open community driven sensing of environment’, Proceedings of the ACM
SIGSPATIAL International Workshop on GeoStreaming, San Jose, CA, pp. 39-42.
Al-Maskari, S.S., Saini, D.K. and Omar, W.M. (2010), ‘Cyber infrastructure and data quality
for environmental pollution control in Oman’, Proceeding of International Conference on
Data Analysis, Data Quality & Metadata Management, Mandarin Orchard, Singapore, doi:
10.5176/978-981-08-6308-1_D-038.
Arasu, A., Babcock, B., Babu, S., Datar, M., Ito, K., Nishizawa, I., et al. (2003), ‘STREAM: the
Stanford stream data manager’, IEEE Data Engineering Bulletin, Vol. 26, No. 1, pp. 19-26.
Bootsma, R.J., Marteniuk, R.G., Mackenzie, C.L. and Zaal, F.T.J.M. (1994), ‘The speedaccuracy trade-off in manual prehension: effects of movement amplitude, object size and
object width on kinematic characteristics’, Experimental Brain Research, Vol. 98, No. 3,
pp. 535-541.
Crumiere, M. (1999), ‘Artificial neural network prediction of ground-level ozone concentration
in Palm Beach City’, Unpublished master thesis, Folrida Atlantic University, Boca Raton, FL.
European Commission. (2001), IST 2001: Technologies Serving People, European Commission,
Rue de la Loi, Belgium.
Fayyad, U., Piatetsky-Shapiro, G. and Smyth, P. (2006), ‘From data mining to knowledge
discovery: an overview’, in Fayyad, U., Piatetsky-Shapiro, G., Smyth, P. and Uthurusamy,
R. (Eds.), Advances in Knowledge Discovery and Data Mining, AAAI/MIT Press, Menlo
Park, CA, pp. 1-34.
Foster, I. (2002), ‘The grid: a new infrastructure for 21st century science’, Physics Today, Vol.
55, No. 2, pp. 42-47.
Friedman, J.H. (n.d.), ‘Data mining and statistics: what’s the connection?’, available at http://
statweb.stanford.edu/~jhf/ftp/dm-stat.pdf (accessed 21 November 2014).
Gehrke, J. and Madden, S. (2004), ‘Query processing in sensor networks’, IEEE Pervasive
Computing, Vol. 3, No. 11, pp. 46-55.
06-Saini.indd 41
2015/4/1 下午 02:07:36
42 Dinesh Kumar Saini, Sanad Al Maskari
Green, G.C., Chan, A.D.C. and Goubran, R.A. (2009), ‘Identification of food spoilage in the
smart home based on neural and fuzzy processing of odour sensor responses’, Proceedings
of Annual International Conference of the IEEE Engineering in Medicine and Biology
Society, Minneapolis, MN, pp. 2625-2628.
Helmbold, D.P. and Long, P.M. (1994), ‘Tracking drifting concepts by minimizing
disagreements’, Machine Learning, Vol. 14, pp. 27-45.
Hill, J., Szewczyk, R., Woo, A., Hollar, S., Culler, D. and Pister, K. (2000), ‘System architecture
directions for network sensors’, ASPLOS, Vol. 35, No. 11, pp. 93-104.
Kahn, J.M., Katz, R.H. and Pister, K.S.J. (1999), ‘Next century challenges: mobile networking
for “smart dust”’, Proceedings of ACM/IEEE International Conference on Mobile
Computing and Networking, Seattle, WA, pp. 271-278.
Li, Y., Callahan, T., Darnell, E., Harr, R., Kurkure, U. and Stockwood, J. (2000), ‘Hardwaresoftware co-design of embedded reconfigurable architectures’, Proceedings of the Design
Automation Conference, Los Angeles, CA, pp. 507-512.
Pan, L. and Yang, S.Y. (2007), ‘A new intelligent electronic nose system for measuring and
analyzing livestock and poultry farm odours’, Environment Monitoring and Assessment,
Vol. 135, pp. 399-408.
Roppel, T.A., Padgetta, M.L., Waldemark, J. and Wilson, D. (1998), ‘Feature-level signal
processing for near-real-time odor identification’, in Dubey, A.C. and Harvey, J.F. (Eds.),
The SPIE Conference on Detection and Remediation Technologies for Mines and Minelike
Targets III, Orlando, FL, pp. 13-17.
Saini, D.K. and Yousif, J.H. (2013), ‘Environmental scrutinizing system based on soft
computing technique’, International Journal of Computer Applications, Vol. 62, No. 13,
pp. 45-50.
Sohn, J.H., Smith, R., Yoong, E., Leis, J.W. and Galvin, G. (2003), ‘Quantification of odours
from piggery effluent ponds using an electronic nose and an artificial neural network’,
Biosystems Engineering, Vol. 86, No. 4, pp. 399-410.
Wood, A. and Stankovic, J.A. (2000), ‘Denial of service in sensor networks’, IEEE Computer,
Vol. 35, No. 10, pp. 54-62.
06-Saini.indd 42
2015/4/1 下午 02:07:36
Data Management Issues and Data Mining of Real Time System Application for Environment Monitoring 43
About the authors
Dinesh Kumar Saini is working as Associate Professor in Faculty of Computing and
Information Technology Sohar University which is affiliated with university of Queensland.
He received his Ph.D. in software Systems, Prior to that he has done his Masters of
Engineering in software systems. He is member of major professional bodies. He has been
actively engaged in teaching and research since last 16 years. His main research interests
are in Environmental Informatics, Learning Content Management Systems, Searching &
Recommending Techniques, Mathematical Modeling, Simulation, Cyber Defense, Network
Security, Computational Intelligent Techniques, Software Testing and Quality.
Corresponding author. Faculty of Computing and Information Technology, Sohar
University, Oman. Research Fellow and Adjunct Faculty, School of ITEE, University
of Queensland, Australia. P.O.Box.-44, PC-311, Sohar, Sultanate of Oma. Tel: (+968)
26720101. E-mail address: [email protected], [email protected]
Sanad Al Maskari is working as Lecturer in the Faculty of Computing and Information
Technology Sohar University. Currently he is on study leave for pursuing his Ph.D. in the
university of Queensland Brisbane Australia. E-mail address: sanad.almaskari@uqconnect.
edu.au
06-Saini.indd 43
2015/4/1 下午 02:07:36
06-Saini.indd 44
2015/4/1 下午 02:07:36